Distance estimation to objects and free-space boundaries in autonomous machine applications
In various examples, a deep neural network (DNN) is trained-using image data alone-to accurately predict distances to objects, obstacles, and/or a detected free-space boundary. The DNN may be trained with ground truth data that is generated using sensor data representative of motion of an ego-vehicl...
Gespeichert in:
Hauptverfasser: | , , , , , , |
---|---|
Format: | Patent |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | |
---|---|
container_issue | |
container_start_page | |
container_title | |
container_volume | |
creator | Oh, Sangmin Park, Minwoo Kwon, Junghyun Yang, Yilin Nister, David Jujjavarapu, Bala Siva Sashank Ye, Zhaoting |
description | In various examples, a deep neural network (DNN) is trained-using image data alone-to accurately predict distances to objects, obstacles, and/or a detected free-space boundary. The DNN may be trained with ground truth data that is generated using sensor data representative of motion of an ego-vehicle and/or sensor data from any number of depth predicting sensors-such as, without limitation, RADAR sensors, LIDAR sensors, and/or SONAR sensors. The DNN may be trained using two or more loss functions each corresponding to a particular portion of the environment that depth is predicted for, such that-in deployment-more accurate depth estimates for objects, obstacles, and/or the detected free-space boundary are computed by the DNN. In some embodiments, a sampling algorithm may be used to sample depth values corresponding to an input resolution of the DNN from a predicted depth map of the DNN at an output resolution of the DNN. |
format | Patent |
fullrecord | <record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_US11769052B2</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>US11769052B2</sourcerecordid><originalsourceid>FETCH-epo_espacenet_US11769052B23</originalsourceid><addsrcrecordid>eNqNzL0KwlAMQOEuDqK-Q3yAgq2ouPqHuzo5lPQ2xUhvcmnS91eKD-B0lo8zzZ4nNkcJBGTOEZ1VwBW0flNwA5QG2p4ot4RfVOsgDfZMBiyAg6to1MEgYnixEGBKHYdxY_Ns0mJntPh1li0v5_vxmlPSisahkFePW1HstvvVpjyU63_MB64RPDA</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>Distance estimation to objects and free-space boundaries in autonomous machine applications</title><source>esp@cenet</source><creator>Oh, Sangmin ; Park, Minwoo ; Kwon, Junghyun ; Yang, Yilin ; Nister, David ; Jujjavarapu, Bala Siva Sashank ; Ye, Zhaoting</creator><creatorcontrib>Oh, Sangmin ; Park, Minwoo ; Kwon, Junghyun ; Yang, Yilin ; Nister, David ; Jujjavarapu, Bala Siva Sashank ; Ye, Zhaoting</creatorcontrib><description>In various examples, a deep neural network (DNN) is trained-using image data alone-to accurately predict distances to objects, obstacles, and/or a detected free-space boundary. The DNN may be trained with ground truth data that is generated using sensor data representative of motion of an ego-vehicle and/or sensor data from any number of depth predicting sensors-such as, without limitation, RADAR sensors, LIDAR sensors, and/or SONAR sensors. The DNN may be trained using two or more loss functions each corresponding to a particular portion of the environment that depth is predicted for, such that-in deployment-more accurate depth estimates for objects, obstacles, and/or the detected free-space boundary are computed by the DNN. In some embodiments, a sampling algorithm may be used to sample depth values corresponding to an input resolution of the DNN from a predicted depth map of the DNN at an output resolution of the DNN.</description><language>eng</language><subject>CALCULATING ; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS ; COMPUTING ; CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE ORDIFFERENT FUNCTION ; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES ; COUNTING ; ELECTRIC DIGITAL DATA PROCESSING ; PERFORMING OPERATIONS ; PHYSICS ; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TOTHE CONTROL OF A PARTICULAR SUB-UNIT ; TRANSPORTING ; VEHICLES IN GENERAL</subject><creationdate>2023</creationdate><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20230926&DB=EPODOC&CC=US&NR=11769052B2$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,308,780,885,25564,76547</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&date=20230926&DB=EPODOC&CC=US&NR=11769052B2$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>Oh, Sangmin</creatorcontrib><creatorcontrib>Park, Minwoo</creatorcontrib><creatorcontrib>Kwon, Junghyun</creatorcontrib><creatorcontrib>Yang, Yilin</creatorcontrib><creatorcontrib>Nister, David</creatorcontrib><creatorcontrib>Jujjavarapu, Bala Siva Sashank</creatorcontrib><creatorcontrib>Ye, Zhaoting</creatorcontrib><title>Distance estimation to objects and free-space boundaries in autonomous machine applications</title><description>In various examples, a deep neural network (DNN) is trained-using image data alone-to accurately predict distances to objects, obstacles, and/or a detected free-space boundary. The DNN may be trained with ground truth data that is generated using sensor data representative of motion of an ego-vehicle and/or sensor data from any number of depth predicting sensors-such as, without limitation, RADAR sensors, LIDAR sensors, and/or SONAR sensors. The DNN may be trained using two or more loss functions each corresponding to a particular portion of the environment that depth is predicted for, such that-in deployment-more accurate depth estimates for objects, obstacles, and/or the detected free-space boundary are computed by the DNN. In some embodiments, a sampling algorithm may be used to sample depth values corresponding to an input resolution of the DNN from a predicted depth map of the DNN at an output resolution of the DNN.</description><subject>CALCULATING</subject><subject>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</subject><subject>COMPUTING</subject><subject>CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE ORDIFFERENT FUNCTION</subject><subject>CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES</subject><subject>COUNTING</subject><subject>ELECTRIC DIGITAL DATA PROCESSING</subject><subject>PERFORMING OPERATIONS</subject><subject>PHYSICS</subject><subject>ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TOTHE CONTROL OF A PARTICULAR SUB-UNIT</subject><subject>TRANSPORTING</subject><subject>VEHICLES IN GENERAL</subject><fulltext>true</fulltext><rsrctype>patent</rsrctype><creationdate>2023</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNqNzL0KwlAMQOEuDqK-Q3yAgq2ouPqHuzo5lPQ2xUhvcmnS91eKD-B0lo8zzZ4nNkcJBGTOEZ1VwBW0flNwA5QG2p4ot4RfVOsgDfZMBiyAg6to1MEgYnixEGBKHYdxY_Ns0mJntPh1li0v5_vxmlPSisahkFePW1HstvvVpjyU63_MB64RPDA</recordid><startdate>20230926</startdate><enddate>20230926</enddate><creator>Oh, Sangmin</creator><creator>Park, Minwoo</creator><creator>Kwon, Junghyun</creator><creator>Yang, Yilin</creator><creator>Nister, David</creator><creator>Jujjavarapu, Bala Siva Sashank</creator><creator>Ye, Zhaoting</creator><scope>EVB</scope></search><sort><creationdate>20230926</creationdate><title>Distance estimation to objects and free-space boundaries in autonomous machine applications</title><author>Oh, Sangmin ; Park, Minwoo ; Kwon, Junghyun ; Yang, Yilin ; Nister, David ; Jujjavarapu, Bala Siva Sashank ; Ye, Zhaoting</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_US11769052B23</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>eng</language><creationdate>2023</creationdate><topic>CALCULATING</topic><topic>COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS</topic><topic>COMPUTING</topic><topic>CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE ORDIFFERENT FUNCTION</topic><topic>CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES</topic><topic>COUNTING</topic><topic>ELECTRIC DIGITAL DATA PROCESSING</topic><topic>PERFORMING OPERATIONS</topic><topic>PHYSICS</topic><topic>ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TOTHE CONTROL OF A PARTICULAR SUB-UNIT</topic><topic>TRANSPORTING</topic><topic>VEHICLES IN GENERAL</topic><toplevel>online_resources</toplevel><creatorcontrib>Oh, Sangmin</creatorcontrib><creatorcontrib>Park, Minwoo</creatorcontrib><creatorcontrib>Kwon, Junghyun</creatorcontrib><creatorcontrib>Yang, Yilin</creatorcontrib><creatorcontrib>Nister, David</creatorcontrib><creatorcontrib>Jujjavarapu, Bala Siva Sashank</creatorcontrib><creatorcontrib>Ye, Zhaoting</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Oh, Sangmin</au><au>Park, Minwoo</au><au>Kwon, Junghyun</au><au>Yang, Yilin</au><au>Nister, David</au><au>Jujjavarapu, Bala Siva Sashank</au><au>Ye, Zhaoting</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>Distance estimation to objects and free-space boundaries in autonomous machine applications</title><date>2023-09-26</date><risdate>2023</risdate><abstract>In various examples, a deep neural network (DNN) is trained-using image data alone-to accurately predict distances to objects, obstacles, and/or a detected free-space boundary. The DNN may be trained with ground truth data that is generated using sensor data representative of motion of an ego-vehicle and/or sensor data from any number of depth predicting sensors-such as, without limitation, RADAR sensors, LIDAR sensors, and/or SONAR sensors. The DNN may be trained using two or more loss functions each corresponding to a particular portion of the environment that depth is predicted for, such that-in deployment-more accurate depth estimates for objects, obstacles, and/or the detected free-space boundary are computed by the DNN. In some embodiments, a sampling algorithm may be used to sample depth values corresponding to an input resolution of the DNN from a predicted depth map of the DNN at an output resolution of the DNN.</abstract><oa>free_for_read</oa></addata></record> |
fulltext | fulltext_linktorsrc |
identifier | |
ispartof | |
issn | |
language | eng |
recordid | cdi_epo_espacenet_US11769052B2 |
source | esp@cenet |
subjects | CALCULATING COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS COMPUTING CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE ORDIFFERENT FUNCTION CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES COUNTING ELECTRIC DIGITAL DATA PROCESSING PERFORMING OPERATIONS PHYSICS ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TOTHE CONTROL OF A PARTICULAR SUB-UNIT TRANSPORTING VEHICLES IN GENERAL |
title | Distance estimation to objects and free-space boundaries in autonomous machine applications |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-24T14%3A34%3A05IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-epo_EVB&rft_val_fmt=info:ofi/fmt:kev:mtx:patent&rft.genre=patent&rft.au=Oh,%20Sangmin&rft.date=2023-09-26&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3EUS11769052B2%3C/epo_EVB%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |